Optical Spectroscopy of Recombining Ions in Flowing
Afterglow Plasmas
R. Johnsen, M. Skrzypkowski, T. Gougousi, R. Rosati, and M.F. Golde
University of Pittsburgh
Key words:
1.
INTRODUCTION
Spectroscopy has been an important tool in studies of dissociative
recombination (DR) for many years. The early work by Biondi1 and coworkers established the dissociative nature of the process by demonstrating
the Doppler broadening of spectral lines due to the momentum that is
imparted on the separating recombination fragments. In more recent work,
spectroscopic methods have been used to identify products of DR of
diatomic and polyatomic ions and to quantitatively measure yields for
specified states of radiating recombination products. The formation of
products in radiating states is of potential interest for the interpretation of
planetary spectra and perhaps those observed from interstellar clouds, and of
course, experimenters hope that their identification of product states will
lead to new insights into the mechanisms of recombination.
Both stationary or flowing afterglow plasmas are excellent media for
studying emissions from recombining molecular ions. A good introduction
and references to the relevant methods can be found in the review article by
Adams2 who carried out a survey of emission spectra from several
recombining ions. Quantitative spectroscopy is considerably more difficult
and hence data on quantitative yields of radiating species are still quite
sparse. As was pointed out by Herbst3 in this meeting, absolute yields are
1
2
R. Johnsen, M. Skrzypkowski, T. Gougousi, R. Rosati, and M.F.
Golde
needed for the interpretation of spectra from interstellar clouds. Such
measurements are being carried in our group at the University of Pittsburgh
and they are the principal subject of this contribution to the symposium.
It is acknowledged that, compared to spectroscopic afterglow methods, ionstorage rings (ISR)4 are a more powerful tool for measuring complete
product yields into all channels, but at this time it is not possible to carry out
optical spectroscopy in ion storage rings due to the low concentration of ions
and electrons. Thus, we view afterglow methods and the ISR techniques as
complementary rather than as competing with each other.
In the following, we will use the term “spectroscopic yield “ to denote the
fraction of recombination events that results in formation of a specified
electronic state. This yield is not necessarily identical to the nascent
recombination yield of this product state since the state may also be
populated by a radiative transition from a higher state that is formed in
recombination. The nascent yield may be obtained by measuring and
subtracting the cascading contribution, if there is any, from higher states.
The procedure usually requires comparing photon count rates that are
observed at widely different wavelength, which in turn requires a precise
knowledge of the wavelength-dependence of the photometric sensitivity of
the optical detection system. The finer points will be discussed later.
2.
STATIONARY AND FLOWING AFTERGLOWS
In a stationary (or static) afterglow system5, a plasma is generated by
exposing a suitable gas mixture, generally a rare gas with small admixtures
of other gases, to a pulsed microwave or other electrical discharge, or to an
intense flux of ultraviolet radiation, or perhaps a high-energy electron beam.
The microwave discharge is used most often. Upon termination of the
discharge, the electron density decay is measured by using Langmuir probes,
by measuring the frequency shift of a resonant microwave cavity mode, or
by measuring the real part of the plasma conductivity using a radiofrequency method. Other diagnostic tools typically include a mass
spectrometer that records the evolution of the relative ion concentrations as a
function of afterglow time.
The spectroscopic measurements most often employ monochromators,
sometimes interference filters, and, for line-shape measurements, FabryPerot interferometers. Photon counting techniques are used almost
exclusively. The analysis of the plasma decay in stationary afterglows is
Optical Spectroscopy of Recombining Ions in Flowing Afterglow
Plasmas
3
simple when only a single-ion species is present and when ambipolar
diffusion is very slow. In general, numerical solutions are required. Since the
intensity of emissions that result from short-lived radiating recombination
products is proportional to the square of the electron density ne, a simple but
very useful test of the consistency of experimental observations consists of
comparing the rates of decay of the electron density, the square root of the
emission intensity, and the mass selected ion currents that are detected by the
mass spectrometer (where available).
The stationary afterglow method has some strength but also some
weaknesses. One problem is that the molecular additives that are required to
make the desired ions are exposed to an intensive discharge, which may
result in dissociation, electronic and vibrational excitation, and formation of
new chemical species. Vibrational excitation of the neutral gases can persist
into the afterglow, produce heating of the electron gas and persistent
vibrational excitation of the ions under study6. In pure rare gases, of course,
such problems do not arise.
The flowing afterglow, to be described in detail later, has the virtue of
allowing far better control over the ion-chemistry than is possible in the
stationary afterglow. Here, the molecular gases are added downstream from
the discharge region and they are never directly exposed to the discharge that
typically operates in pure helium. Energetic particles, mainly metastable
helium atoms, are safely converted to argon ions before molecular gases are
added.
A brief note of warning to flowing-afterglow users: While helium
metastables are indeed removed by Penning ionization with argon, the
resulting atomic argon ions recombine with electrons by three-body
recombination and form metastable argon atoms (with concentrations of 10
to 20 % of the ion density) and those can lead to unexpected reactions. Also,
there is some evidence that radiatively trapped ultraviolet photons from the
discharge enter the reaction region. This seems to happen in spite of the
usual precaution of putting a right-angle bend into the discharge inlet such
that there is no line-of-sight path to the reaction region.
Modeling of the ion-kinetic processes in flowing afterglow plasmas is almost
always required. In the work in Pittsburgh, we use computer models that
include the relevant kinetic processes, ambipolar diffusion in gas flowing
with the proper parabolic dependence of flow velocity on the radial
coordinate, and an approximate treatment of the gas mixing downstream
from the gas injection port.
4
2.1
R. Johnsen, M. Skrzypkowski, T. Gougousi, R. Rosati, and M.F.
Golde
Stationary-afterglow data on molecular dimer ions
We mention this subject here, in part, to draw attention to the problem of
accounting for the anomalously large recombination coefficients of cluster
and dimer ions. It is striking that weakly-bound molecular dimer ions, such
as N2+N2 and CO+CO recombine much faster than the monomers. At one
time D. Bates7 postulated that such ions undergo what he called “super
dissociative recombination” and he ascribed the high recombination rate to
the availability of dissociation channels into excited states of the molecular
fragments. Bates proposed that the super-dissociative mechanism is also
responsible for the high recombination rates of the weakly-bound rare-gas
dimer ions (except for He2+) which were the first ions studied by Biondi1
Partly in response to Bates’ suggestion, Cao and Johnsen8 carried out
measurements on N2+N2 recombination in a high-pressure (several hundred
Torr) stationary afterglow system that was well-suited for this work because
the high pressure promotes rapid conversion of the monomer to dimers ions.
In addition, COCO+ ions were studied (unpublished). The recombination
rates indeed were found to be very large (~2.5 x 10-6 cm3/s) and both
exhibited strong emissions of molecular bands. In the case of N4+, the
second positive band system of N2 (C3u – B3g) was observed while three
band systems were found to arise from recombination of COCO+, the third
positive band (b3 –a3), the Angström band ( B1-A1), and the Herman
band (e3-a3) . The origin of the emissions was clearly recombination, as
evidenced by their disappearance on adding an electron scavenger (SF6) and
the observed scaling of intensities with ne2.
The observation of the molecular bands supports Bates conjecture, but since
absolute yields were not determined it has not been shown that the observed
dissociation channels are dominant and responsible for the large
recombination coefficient. The same statement applies to the recombination
of rare-gas dimer ions; we still do not know if the spectroscopically observed
states are indeed dominant. It would be interesting to perform measurements
on dimer ions in an ion storage ring experiment to see if (a) similar high rate
coefficients are found and (b) if other product channels exist.
Optical Spectroscopy of Recombining Ions in Flowing Afterglow
Plasmas
3.
5
ABSOLUTE SPECTROSCOPY IN FLOWING
AFTERGLOWS
We now turn our attention to measurements of absolute yields. The flowingafterglow system9 that we used has the configuration shown in Fig. 1.
Metastable helium atoms, produced in a microwave discharge in helium, are
converted by Penning ionization to Ar+ ions by downstream addition of
argon (density ~ 15% of that of helium). Ion and electron densities at this
point are ~ 4x1010 cm-3. The reagent gases that initiate formation of the
desired molecular ions are added through movable and fixed reagent inlets.
The diagnostic tools include a quadrupole mass filter that analyzes the ion
composition of the plasma at a point downstream from the reagent inlet, and
a movable Langmuir probe (diameter = 25m, length = 3-4 mm) that
measures electron densities. The Langmuir probes are used in the electroncollecting mode since a series of tests10 showed that ion-collecting probes
produce misleading results.
Emission spectroscopy of the plasma is carried out by a movable (along the
direction of the flow) monochromator/phototube system (McPherson 0.3 m
monochromator, Hamamatsu R943-02 phototube) that views the plasma
through a long quartz window. The system covers the spectral range from
200 to 900 nm. An in-situ absolute calibration of the optical systems was
performed by measuring the intensity of the N2(C3u-B3g)(0,0) band at 337
nm, excited by the reaction
Ar* + N2  Ar + N2(C3u)
(yield 0.800.20) )11,12
The flux of Ar* was determined by converting the Ar* atoms to NO+ ions in
the reaction
Ar* + NO  Ar + NO+ + e (yield 0.350.10)13,14
and using the Langmuir probe to measure the electron density increase due
to this reaction. The spectral response of the system at other wavelengths
was measured using a calibrated tungsten lamp in the red, a deuterium lamp
in the uv-to visible range, and joining the two response curves in the
overlapping spectral range.
To distinguish emissions from recombining ions from those of extraneous
sources, an electron scavenging gas (SF6) was added.
6
R. Johnsen, M. Skrzypkowski, T. Gougousi, R. Rosati, and M.F.
Golde
Figure 1. Schematic diagram of the flowing-afterglow tube
The methods of analyzing flow-tube data are very well known and will not
be described here. We use computer models that allow accurate modeling of
diffusion, gas mixing, ion-molecule reactions, recombination, and product
formation. Some details of these methods have been presented in a recent
publication 9.
3.1
Emission spectra and yields from DR of N2O+,
HCO2+, N2OH+, HCO+/COH+, and CO2+
In this section, we will give a few examples of recent and ongoing flowingafterglow measurements of spectroscopic yields from the recombination of
several ions. Due to space limitations, we cannot reproduce figures of all
spectral scans. Some may be found in a recent book chapter15.
Many recombining ion species were found to produce measurable spectral
emissions but the absolute yields of the radiating states were often found to
be quite small.
Optical Spectroscopy of Recombining Ions in Flowing Afterglow
Plasmas
7
For instance, the yield for NO(B2i) from the recombination
N2O+ + e 
NO(B2i) + N(4S),
as determined from the intensity of the NO (B2-X 2band, is only 0.6
percent. The spectra show additional weak emissions of the N2 (C3 - B3)
bands and the NO(A2+- X2) bands.
Similarly, the yield of the recombination branch
HCO2+ + e 
OH(A2+) + CO(X1+) ,
as determined from the intensity of the observed OH(A2+-X2i) (v=0 and
v=1) band system, also is quite small (1.5%).
One interesting result of this work, which still awaits a complete analysis, is
the finding that both NH(A3i-X3) and OH(A2+-X2i) spectra are
emitted in recombining plasmas containing N2OH+ ions. This suggests that
the emissions arise from recombination of the two isomeric species HNNO+
and NNOH+
NNOH+ + e 
HNNO+ + e 
OH(A2+) + N2(X1g+)
NH(A3i) + NO(X2r)
(2 %)
(1-5 %)
The intensity of the NH band decayed far more rapidly with axial position
than that of OH. This observation supports the assumptions that the bands
arise from two distinct ions species, while ruling out the alternative
explanation that recombination involves a major rearrangement of the
molecule. The yields of both radiating states are again quite small, 2% for
OH(A2+) and 1-5 % for NH(A3i).
Recombination of HCO+ and its isomeric form COH+ are interesting
processes for astrophysical applications. In a still ongoing experiment, we
study the emission spectra that arise from the dissociative recombination of
HCO+/ COH+ by producing a mixture of those ions using the a reaction
scheme that starts from
Ar+ + H2  ArH++ H
and
ArH+ + H2  H3+ + H.
By adding CO and H2 the subsequent reactions16
8
R. Johnsen, M. Skrzypkowski, T. Gougousi, R. Rosati, and M.F.
Golde
H3+ + CO  HCO+ + H2
 COH+ +H2
(94%)
( 6%)
and the reactions
ArH+ + CO  Ar + HCO+
 Ar + COH+
Ar+ + CO  Ar + CO+
and
CO+ + H2  HCO+ +H
 COH+ +H
( ~50%)
(~50%)
produce both isomers17. Kinetic modeling, based on currently available
branching fractions, shows that the fractional concentration of the more
energetic COH+ isomer should remain small (several %) under our
conditions. Conversion of COH+ to HCO+ occurs by the reactions18
and
COH+ + CO  HCO+ + CO
COH+ +H2  HCO+ + H2
The intensity of the wavelength-integrated uv-spectra show that
recombination of HCO+/ COH+ results in CO(a3r) with an effective yield
(including cascading from higher states) of (26+ 10) % for a mixture of the
two isomeric forms. Since HCO+ is far more abundant, the CO(a3r) product
must predominantly arise from HCO+. However, several intense visible
emissions (e-a, d-a, a’-a) from higher-lying states of CO are observed also. It
is energetically possible to populate the e and d states in recombination of
COH+ but not in recombination of HCO+.
Optical Spectroscopy of Recombining Ions in Flowing Afterglow
Plasmas
9
Figure 2. Spectral scan from 400 to 700 nm of (e-a, d-a, a’-a) band emissions from
recombination of a mixture of HCO+/COH+ ions. The atomic helium and argon lines are due
to scattered light from the discharge and upstream region of the plasma.
Fig 2 shows a spectrum of the overlapping visible emissions from the
higher-lying states of CO. Clearly, the spectrum contains much information
but it is difficult to identify the origin of particular features. It helps to
investigate how the spectrum changes with the axial position in the flow tube
(‘afterglow time’). Measurements show that emissions from the d state and
from a’(v>4) levels decay much faster than those originating from the a’
(v<4) levels. This seems to indicate that the d state is populated by
recombination of COH+ while different vibrational levels of a’ may be
populated by recombination of HCO+ and COH+. A faster decay of COH+ is
to be expected since this isomer is destroyed (converted to HCO+) in the
reactions with H2 and CO. The data available at the time of this writing are
compatible with the assumption that both isomers recombine with the same
rate, but more data need at different conditions have to be taken to reach a
firm conclusion.
10
R. Johnsen, M. Skrzypkowski, T. Gougousi, R. Rosati, and M.F.
Golde
Our earlier published work on CO2+ recombination19 indicated substantial
emissions from the long-lived a3r state. This confirmed the work by
Wauchop and Broida,20 and Taieb and Broida 21 . It was concluded that this
state is populated directly by the recombination channel
CO2+ + e

CO(a3r) + O(3P)
and, in addition, by cascading from higher states (a'3+, d 3i, e3) but the
magnitude of cascading contribution was not determined at that time. The
spectroscopic yield (~29%) of CO(a3r) was only about one half of that
obtained in the earlier work(55 %) 20, 21. After a more thorough calibration of
the optical system and additional model calculations, we concluded that
cascading contributes from 50 to 100 % to the long-lived CO(a). The
combined yield for the states a’, d, and e was found to be (22+7) %. Using
the relative yields measured by Tsuji et al 22 , the absolute yields would be
13 % for a’, 8% for d, and 1.7% for e. The uncertainty in the cascading
contribution to CO(a) is still rather large. One problem is that our computer
models do not accurately reproduce the z-dependence of the measured
intensities in the visible range, especially in the vicinity of the reagent inlet.
Recent modifications of the gas inlet and a reduction of light scattering in the
flow tube have improved the situation and we hope to be able to arrive at a
more accurate determination of the cascading contributions.
4.
NO EMISSIONS FROM DR OF H3+?
We had expected to observe emissions from H3+ recombination since some
experimental work suggested that such emissions should be present (see also
discussion and references in Gougousi et al.23). The very recent work by
Glosik and coworkers in Prague 24 strongly suggests that the recombination
of H3+ proceeds via a long-lived H3 complex which can reasonably be
expected to undergo radiative transitions, as had been suggested earlier by
Gougousi et al. However, no spectral emissions that could be ascribed to
dissociative recombination of H3+ were found by us in the wavelength range
from 200 to 900. Adams25 reported a similar negative finding. It seems
ironic that the first observations of the H3 spectrum by Herzberg and
coworkers (see e.g. Dabrowski and Herzberg26) were ascribed to
recombination. At this time, the most likely explanation of the origin of the
spectra observed by Herzberg and coworkers may be that the process
Optical Spectroscopy of Recombining Ions in Flowing Afterglow
Plasmas
11
responsible was dissociative recombination of H5+ ions rather than of H3+. In
this context, the observation of line broadening in some H3 rotational lines
by Miderski and Gellene 27 is very interesting. The authors ascribed the
broadening to the energy released in dissociative recombination of H 5+ into
H2 and H3* . It would be interesting to repeat these experiments in a cooled
flowing afterglow experiment and with higher-resolution interferometric
techniques.
5.
CONCLUSIONS
The spectroscopy of recombining afterglow plasmas provides a rich and
important tool for investigating dissociative recombination process.
However, the interpretation of observed spectra is subtle and complicated,
more so if one seeks to obtain absolute yields for particular states. In many
cases, the yields for radiating states are small. That does not make them
uninteresting since the observed spectral features may still be useful for
diagnostic purposes in laboratory and space plasmas. One also quickly
discovers that the recombining plasma may contain mixtures of isomeric
form of the same ion and that those can give rise to different spectra. While
this complicates the interpretation further, it also allows studies of the
different recombination behavior of different isomeric forms, which are
difficult to do by other methods.
The recombination mechanism of H3+ ions received much attention during
this symposium but it remains an ‘enigma’. The absence of measurable
emissions may contain some hints but we do not know in which direction
they point.
ACKNOWLEDGMENT
This work was, in part, supported by the NASA Planetary Atmospheres Program
1
2
3
4
5
M.A. Biondi, elsewhere in this volume
N.G. Adams, "Spectroscopic Determination of the Products of Electron-Ion
Recombination. In "Advances in Gas Phase Ion Chemistry", Volume 1, pages 271-310,
JAI Press 1992
E. Herbst, elsewhere in this volume
S. Datz, G. Sundström, C. Biederman, L. Broström, H. Danared, Phys. Rev. Lett.
77,4891(1995)
see also M.A. Biondi, elsewhere in this volume
12
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
R. Johnsen, M. Skrzypkowski, T. Gougousi, R. Rosati, and M.F.
Golde
R. Johnsen, Int. J. Mass Spectr and Ion Proc. 81, 67 (1987)
D.R. Bates, J.Phys. B 24, 703 (1991)
Y.S. Cao and R. Johnsen, J. Chem. Phys. 95, 7356 (1991) and Y.S. Cao, PhD thesis,
University of Pittsburgh (1991)
T. Gougousi, R. Johnsen, M. F. Golde, Journal of Chemical Physics 107, 2430 (1997)
R. Johnsen, E.V. Shun'ko, T. Gougousi, and M.F. Golde, Phys. Rev E 50, 3994 (1994)
J. H. Kolts, H. C. Brashears, and D. W. Setser, J. Chem. Phys. 67, 2931 (1977)
N. Sadeghi and D. W. Setser, Chem. Phys. Lett. 82, 44 (1981)
M. F. Golde, Y.-S. Ho, and H. Ogura, J. Chem. Phys. 76, 3535 (1982)
M. T. Jones, T. D. Dreiling, D. W. Setser, and R. N. McDonald, J. Phys. Chem. 89, 4501
(1985).
R. Johnsen, M. Skrzypkowsi, T. Gougousi, and M.F. Golde, in Dissociative
Recombination, Theory, Experiment and Applications (World Scientific 2000, M. Larrson,
J.B. Mitchell, I.F. Schneider eds)
A.J. Illies, M.F. Jarrold,, and M.T. Bowers, J. Am. Chem. Soc. 105, 2562, 1982)
C.G. Freeman, J.S. Knight, J.G. Love,, and M.J. McEwan, Int. J. Mass Spectrom and Ion
Processes 80, 255, (1987)
C.G. Freeman, J.S. Knight, J.G. Love, M.J. McEwan, Int. J. Mass Spec. and Ion Proc. 80,
255 (1987)
M.P. Skrzypkowski, Theodosia Gougousi, Rainer Johnsen, Michael F. Golde, Journal of
Chemical Physics , J. Chem. Phys. 108, 8400 (1998)
T.S. Wauchop and H.P. Broida, J. Chem. Phys. 56, 330 (1972)
G. Taieb and H.P Broida, Chem. Phys. 21, 313 (1977)
M. Tsuji, K. Kobarai, H. Kouno, H. Obase, and Y. Nishimura, J. Chem. Phys. 94, 1127
(1991)
T. Gougousi, R. Johnsen, and M.F. Golde, Int. J. Mass Spec. and Ion Proc.
149/150, 131 (1995)
J. Glosik, elsewhere in this volume
N.G. Adams, private communication
I. Dabrowski and G. Herzberg, Can. J. Phys. 58, 1238 (1980)
C.A. Miderski and G.I. Gellene, J.Chem. Phys. 88, 5331 (1988)